| Summary: | 碩士 === 國立臺北科技大學 === 工業工程與管理系 === 107 === In recent years, due to the slowdown in the growth of notebook computers and tablet PCs, the performance of major brands has fallen into a bottleneck in research and developments. Since notebook computers are still high-priced products, and products become more sophisticated than desktop computers. In addition, the differences in products make assembly and maintenance to different degrees of difficulty, leading to an extremely high competition in the notebook market. In the past decade, notebook computer repair components often suffered from out of stock or uneven distribution, resulting in significant cost increases for major companies in spare-parts preparation. Until recently, with the rapid developments and remarkable breakthrough of machine learning technology, various very-large scale corporate decision-making problems can be dealt with without undue difficulty.
In this thesis, we will use the extreme gradient boosting method to propose a new prediction model for the demand of important components of the notebook computer for maintenance/repair, which can also be used to trace down the demand tendency during the future warranty period of important components for outdated products. Namely, demand forecasting for important components of notebook manufacturers with an extension to over 12 months is the major objective of this research. By doing so, the personnel of the procurement department is able to take appropriate decisions based on the prediction results in an attempt to provide consumers with prompt and high-quality after-sales service.
|
|---|
